PROJECT SUMMARY Autism Spectrum Disorder (ASD) is a common, costly, heterogeneous neurodevelopmental disorder comprised of impairments related to social communication and interaction as well as restricted interests and repetitive behaviors. Social communication deficits are core to ASD with challenges in this domain often translating into serious lifespan impairment for most individuals. While certain behavioral and pharmacological interventions have some benefits to many children with ASD, such interventions often require significant time and effort- intensive burdens for implementation, suffer from low-adherence and limited availability in community settings, and ultimately have demonstrated weak transfer of skills to real world settings of meaning. While advances in technological capacity have the potential to yield intervention platforms of meaning, current technological systems have focused on simple discrete social skill development or required confederate or remote operation, limits which have impeded progress toward successful mainstream clinical use and benefit. In order to address these issues we propose to develop and test the feasibility, tolerability, and potential for clinical benefit of a novel technological paradigm for meaningful social communication that combines two innovative technologies: (1) collaborative virtual environments (CVE) and (2) artificially intelligent (AI) agent-based interactions to create a virtual intelligent system (VIS) for ASD intervention. We will design a VIS in which children with ASD will participate in a series of engaging game-like interactive tasks where they collaborate with peers and with artificially intelligent partners (AIP). We hypothesize that the proposed paradigm will preserve all advantages of a traditional CVE for meaningful interaction, but will allow unrestricted verbal communication with real partners and eliminate the need for manual coding by providing consistent unbiased measurements of meaningful aspects of social interaction. The proposed system will autonomously index both traditional markers of performance as well as social communicative behaviors contributing to task performance and attempt to modify tasks to meaningfully alter subsequent interactions. It is hypothesized that this measurement and intervention strategy, deployed across engaging, dynamic interactions highly relevant to real-world social interaction, will readily yield quantifiable metrics of important components of social communication that are potentially sensitive to change and responsive to intervention strategies. The specific aims of this project are: 1) Development of a virtual intelligent system and embedded artificially intelligent partners capable of dynamic meaningful interaction during collaborative game activities, and 2) Validation of feasibility and tolerability of the system with children and adolescents with ASD. The proposed technology is simple, cost-effective and is easy to translate for mainstream use.